A Case Study of the Effects of Management Interventions on the Phosphorus Dynamics at a Coastal, Eutrophic, Caribbean lagoon (Laguna Cartagena, Puerto Rico)

: Laguna Cartagena is a coastal, eutrophic, shallow lake and freshwater wetland in southwestern Puerto Rico, managed by the US Fish and Wildlife Service. This ecosystem has been impacted by phosphorus loading from adjacent agricultural areas since the 1950s, causing eutrophication and deteriorating wildlife habitats. Herein, we describe phosphorus input and export during September 2010–September 2011 (Phase One) and October 2013–November 2014 (Phase Two). These two phases bracket a period of intensiﬁed management interventions including excavation and removal of sediment and vegetation, draining, and burning during the summers of 2012 and 2013. Results indicate that Laguna Cartagena retains a phosphorus (sink) in its sediments, and exhibits nutrient-releasing events (source, mainly total phosphorus) to the lagoon water column, which are associated with rainfall and rising water levels. External factors including water level ﬂuctuations and rainfall inﬂuenced phosphorus export during Phase One, but after management interventions (Phase Two), internal processes inﬂuenced sink/source dynamics, releasing elevated phosphorus concentrations to the water column. When exposed sediments were re-ﬂooded, phosphorus concentrations to the water column increased, releasing elevated P concentrations downstream to an estuarine wetlands area and the Caribbean Sea. Herein we offer management recommendations to optimize wildlife habitat without elevating phosphorus concentrations.


Introduction
Phosphorus (P) and nitrogen (N) eutrophication of aquatic ecosystems and a general deterioration in water quality have increased with human population growth and agricultural development within many watersheds, and sources of watershed nutrient loading include both domestic and agricultural effluents [1][2][3][4][5]. Such activities create negative impacts in natural wetland habitats by changing vegetation, soil biogeochemistry, habitat fragmentation, release of greenhouse gases, and severe vegetative overgrowth, resulting in deteriorating wildlife habitats. At our Laguna Cartagena study site, for example, severe cattail overgrowth and extensive peat formation due to phosphorus loading have reduced open surface conditions to as little as 10% (90% cover) and produced poorly oxygenated water [1,6,7].
Excessive nutrient enrichment of P and N to aquatic ecosystems also leads to: excessive growth of nuisance algae, cyanobacteria, and other aquatic plants; high decomposition rates of the accumulated plant biomass that increase the production of organic matter and affect the biogeochemical cycles of N and P; water acidification; decreases in dissolved oxygen (anoxia and hypoxia); and decreases in biological diversity [1,8,9].
P Sink dynamics are defined as occurring when inflow (inlet) P concentrations exceed outflow (outlet) concentrations (the lagoon absorbs P and sends water of lower P concentration downstream to other aquatic environments) [3]. Conversely, P Source dynamics (or internal P loading) occur when outlet P concentrations exceed inlet concentrations (the lagoon releases P and sends water of higher P concentration downstream to other aquatic environments) [3]. Internal P loading can occur from microbial-driven P release and release of inorganic P from mineral binding sites, especially under anaerobic conditions. These processes are controlled by oxygen concentrations, redox conditions, pH, temperature, and other drivers influenced by changes in hydrology [19].
Hydrological and hydroperiod regimes can directly modify and change the physicochemical environment, stimulate microbial activity, and influence nutrient retention (sink) and release (source) processes. These regimes also transform wetland soils (sediments and organic matter) [3,20] and cause eutrophication [7] as sediments and organic matter provide internal sources of nutrients to the water column. At Laguna Cartagena this has resulted in a proliferation of emergent vegetation and, most notably, extensive floating peat mats (islands) and associated floating vegetation. This emergent vegetation covers ca. 90% of the lagoon's surface, blocking light penetration and thus prohibiting the growth and photosynthesis of submerged plants, microalgae, and cyanobacteria. The resulting low dissolved oxygen concentrations negatively impact fisheries, and the overgrowth of floating peat islands reduce the available habitat for waterfowl that require open surface conditions.
Symptoms of severe problems of eutrophication in this shallow lagoon (Laguna Cartagena) have been observed since the 1950s, when most of the land area that drains to the lagoon was used for sugar cane cultivation. This formally oligotrophic freshwater wetland was thus affected by very intense phosphorus loading from inorganic agricultural fertilizer until the end of sugar cane cultivation and subsidized fertilizer use in the late 1990s. A legacy of this fertilizer use may remain in the soils but has not been adequately quantified.
In April 2012, the US Fish and Wildlife Service (USFWS) began a conservation plan that emphasized wildlife diversity, focusing especially on providing habitat for water birds and fish, and thus requiring an open water area of about 70% (30% cover), and welloxygenated water [21]. Hoping to achieve these goals, the agency engaged in management interventions that included the lowering of water levels, completely draining the lagoon by the end of May 2012, then excavating and dredging bottom sediments and removing floating peat mats. The lagoon was subsequently refilled in September of 2012. The lagoon was drained again during July 2013. On 21 August 2013, a prescribed burn of cattaildominated habitat was conducted at Laguna Cartagena as part of a federal government training exercise for federal firefighters and other emergency personnel (as stated in the press). Refilling of the lagoon commenced immediately after the burn, yet within four months (December 2013) the emergent cattails and floating peat mats had returned to their previous extents.
Given the above events, and knowing beforehand the planning of the management interventions, the primary objective of this study was to assess the influence of these management interventions on phosphorus concentrations of the water released by the lagoon downstream in relation to the water that enters the lagoon, by characterizing soluble reactive phosphorus (SRP) and total phosphorus (TP) concentrations entering and leaving the lagoon prior to interventions (Phase One, September 2010 to September 2011) and after interventions (Phase Two, October 2013 to November 2014). A second objective was to determine the role of precipitations and water level changes on P sink and source dynamics.

Description of the Study Site
Laguna Cartagena is situated in the Lajas Valley, Municipality of Lajas, in southwestern Puerto Rico (Figure 1). During the 1950s, the Principal Drainage Canal of the Lajas Valley Irrigation System, adjacent to the town of Magüayo, was connected through Laguna Cartagena. This canal carried agricultural effluent to the lagoon, and the water passed through the lagoon and subsequently flowed westward from the lagoon via an outlet canal ( Figure 1) to the wetlands and estuarine areas of the Boquerón State Bird Refuge, and then to the Caribbean Sea. During 1995 to 1998, SRP (orthophosphate or inorganic P, SRP) concentrations in the lagoon waters ranged to above 1000 and 1500 µg/L [22]. Severe overgrowth of cattails (Typha domingensis) and extensive floating peat mats and floating vegetation still occupy large portions of the available volume and surface area of the lagoon, altering the habitat and thus degrading the abundance and diversity of resident and migratory aquatic birds and fish. The lagoon exhibits other aquatic macrophytes including floating forms such as Pistia stratiotes (water lettuce) and Eichhornia crassipies (water hyacinth) that together with the floating peat mats cover 80-90% of the lagoon surface.

Sampling and Processing
Water samples were collected in triplicate at mid-column depths from the lagoon's single inlet and the lagoon's single outlet to characterize average total P and SRP during the annual cycle. All of the nutrients that enter the lagoon from this half of the drainage basin come together and enter through the combined inlet canal. Sampling occurred at two-week intervals from September 2010 through September 2011 (Phase One of this study) [23][24][25]), prior to management interventions and from October 2013 through November 2014 (Phase Two of this study) [24,25]), after management interventions.
All samples for P (SRP and TP) analyses were collected in glass bottles, preserved with H 2 SO 4, and transported in ice to the Universidad del Turabo (now Ana G. Méndez University, Gurabo). The samples were stored at a low temperature and analyzed within an interval of 1 week and a half after being collected using persulfate digestion and ascorbic acid methods [26]. A Thermo Fisher Scientific, Genesys 20, with a wavelength range between 352 to 1100 nm was used to measure sample absorption at 880 nm.
Rainfall data were obtained weekly from a rain gauge located approximately 6 km north of the lagoon in the upper portion of the watershed, representative of regional rainfall. Daily rainfall data were recorded from August 2010 to September 2011 and September 2013 to November 2014. Stage (water level) data were collected at the adjustable crest weir at the lagoon. Stage was recorded weekly, and for each sampling day between July 2010 to October 2011 and July 2013 to November 2014 to document water level fluctuations. The bottom of the lagoon (0% fill) lies at the 9 m contour above sea level while full stage (100% fill) occurs at the 11 m contour (USGS 7.5-min topographic map), for a potential maximum lagoon depth (100% fill and overflow) of 2 m. Water level is controlled by USFWS's placement and removal of flashboards at the outlet control structure (a weir) and thus are largely independent of rainfall.
We also analyzed samples of consolidated bottom substrate and flocculence in the center of the lagoon and collected general water quality data such as dissolved oxygen (DO), pH, total dissolved solids (TDS), temperature, turbidity, and nitrogen.
Dissolved oxygen, pH, TDS, temperature, and turbidity were measured in the field for each sampling event. The instruments used for measurement of these parameters were the Hanna Instruments (HI 98186) for DO, AZ Instruments Corp, model 8685 for pH and temperature, an EC/TDS/ • C Martini (EC 59) meter for TDS and a turbidity tube for turbidity. Nitrate (NO − 3 ), nitrite (NO − 2 ), and ammonia (NH 3 ) were determined using a HACH DR900 colorimeter. Nitrate was measured using the cadmium reduction method 353.2, nitrite was measured with the USEPA diazotization method and ammonium using the salicylate method. The sediment and flocculence samples were collected in 9 April 2011 (Phase One) and 10 August 2014 (Phase Two). Samples were transported in bags and processed by the Central Analytical Laboratory of the University of Puerto Rico's Agricultural Experimental Station.

Statistical Analyses
Mean SRP and TP values were calculated for each sample (triplicate sampling, n = 3 for each sample). Descriptive statistics are herein provided for both Phase One (2010-2011) and Phase Two (2013-2014). One-way ANOVA and paired t-tests (µ1 > µ2) were used to compare phosphorus concentrations between both phases (Phase One and Phase Two). A multiple regression model was used to determine the response of phosphorus (SRP OI and TP OI ) to several predictors including water level fluctuations (water levels one week prior to water sampling) and rainfall (one or two weeks prior to water sampling) during both phases. To describe the differences between outlet and inlet P concentrations, SRP OI was defined as the subtraction of SRP Outlet − SRP Inlet and TP OI was defined as the subtraction of TP Outlet − TP Inlet .
This study was conducted under the authorization of Biosafety Committee (IBC) Protocol B03-017-13 of the Ana G Mendez University System Office of Compliance.

Rainfall Conditions
Monthly rainfall (mm) for the Laguna Cartagena drainage basin recorded from August 2010 to September 2011 ( Figure 2) and September 2013 to November 2014 ( Figure 3) indicated heavy rainfall during the months of August to November and April to May, and allowed identification of specific rainfall events that occurred one and two weeks prior to water sample collection.

Stage Fluctuations
The maximum depth of 2 m (100% fill and overflow) at Laguna Cartagena was obtained during Phase One on the following dates: 5 October, 9 October, 7 November

Dissolved Oxygen, pH, Total Dissolved Solids, Temperature, Turbidity, and Nitrogen
Overall, Laguna Cartagena had a neutral pH (6.9-7.2) and temperatures that ranged from 25.5 • C to 27.6 • C (Table 1). During Phase Two, DO increased for both the inlet (from 0.34 mg/L to 1.79 mg/L) and lagoon's outlet (from 0.43 mg/L to 3.59 mg/L) ( Table 1). The overall DO concentration for 2013-2014 (Phase Two) was significantly greater than for 2010-2011 (Phase One) ( Figure 5). During both Phase One (33.6 cm) and Phase Two (29.5 cm), the lagoon's outlet had a higher turbidity than the inlet (Table 1).

Analysis from Consolidated Bottom Substrate and Flocculence Associated with Floating Islands
Consolidated bottom substrate and the flocculence associated with floating islands in the center of the lagoon were acidic and with very high nutrient concentrations for both Phase One and Phase Two of this study (Table 2). For Phase One, the pH for consolidated bottom substrate was typically about 4.16 and it was 3.86 for Phase Two. During 2014 the pH for flocculence in the center cattails was higher (4.30) than during 2011 (2.82).

Source/Sink Dynamics under High Versus Low Rainfall and Water Level Fluctuations
During Phase One of this study (pre intervention, 2010-2011), stage fluctuations (water level (WL) one-week prior to sampling) and rainfall (one week (1R) and two weeks (2R) prior to sampling) affected the source-sink dynamics. The data had a normal distribution and the results of the regression equations were, for SRP OI      Overall population mean concentrations of outlet TP, for 2013-2014 (1573.0 µg/L) were significantly greater than for 2010-2011 (703.1 µg/L) using one-way ANOVA (p = 0.011) and paired t-tests (µ 1 > µ 2 ) (p = 0.025) ( Figure 11). Moreover, comparisons of similar sampling dates (Figure 9, Figure 11) show dramatically higher concentrations of TP in Phase Two. Thus, while the phosphorus concentrations of inputs into the lagoon were similar in Phase One and Phase Two, the concentrations of phosphorus in water flowing out of the lagoon were notably higher in the post intervention Phase Two than prior to intervention (Phase One).

Phosphorus Inputs and Outlets Before and After Intervention Measures
The channelized external sources of nutrients that enter Laguna Cartagena through the combined inlet at Magüayo collect the nonpoint sources without any treatment measures. Currently, the major external sources of water and nutrients to the lagoon come together in an inlet canal ("a" in Figure 1) and include runoff and nonpoint nutrient pollution from fertilizers applied to adjacent rice and pineapple fields as well as high density livestock (dairy cattle) pastures (and possibly legacy nutrients in the soils of these areas) on the lands formerly used for sugar cane, and water flows from the lagoon outlet to the downstream habitats mentioned previously.
We found no significant differences between P (SRP and TP) concentrations at the inlet canal for Phase One (2010-2011) versus Phase Two (2013-2014). During Phase One (2010-2011), overall mean concentrations of SRP and TP at the inlet canal were 180.8 µg/L and 840.8.0 µg/L, respectively (Figure 8, Figure 10). During Phase Two (2013-2014), the overall mean concentrations of SRP and TP were 231.5 µg/L and 966.9 µg/L, respectively. At the lagoon's outlet, the mean concentration of TP was significantly higher in Phase Two (1573.0 µg/L) than in Phase One (703.1 µg/L) (Figure 9, Figure 11), and similarly, SRP concentrations at the lagoon's outlet during Phase Two (269.8 µg/L) were substantially higher than those of Phase One (127.7 µg/L) (Figure 9, Figure 11).

Role of Precipitation and Water Level Changes in P Sink and Source Dynamics
During both Phase One (2010-2011) and Phase Two (2013-2014) Laguna Cartagena acted at various times as both a sink and source of TP and SRP. The water column in the lagoon is nitrogen depleted, with very low N:P ratios ( Figure 6). Overall, high nutrient concentrations are stored in the acidic sediment/substrate of the lagoon and management interventions (primarily draining and dredging) occurred in 2012 and 2013. This study addressed two important Laguna Cartagena substrate types: (a) consolidated bottom substrate and (b) flocculence associated with floating islands. The consolidated bottom substrate is compacted clay and the flocculence or organic substrates are produced from the remains of decomposing macrophytes and cattails accumulated in the lagoon. As organic matter decomposes by the action of soil microorganisms, CO 2 is released and as a consequence, the soil is acidified. The plants were growing in matts of floating peat, and water column phosphorus concentrations were so high, especially in relation to the depleted N concentrations, that they were unaffected by plant growth, either before or after the interventions. This is especially evidenced by the fact that upon refilling of the lagoon (after vegetation removal, dredging and draining) the rapidly recovering vegetation did not produce a decline in P concentration from October 2013 to November 2014.
During Phase One (2010-2011), P dynamics within the lagoon were primarily influenced by factors extrinsic to the lagoon: rainfall (one and two weeks prior to sample collection) and water flows to the lagoon (all of which are beyond the control of the management agency). Continual fluctuations in water levels and resulting alternate flooded and drained conditions may substantially alter the stability and redistribution of stored materials including phosphorus (P) in nutrient-impacted wetlands [27,28]. Overall, Laguna Cartagena acted as a sink of SRP and TP under stable conditions, when the inlet concentrations consistently exceeded outlet concentrations, indicating that this ecosystem was assimilating P. Phosphorus that enters Laguna Cartagena tends to accumulate in the sediments and flocculence, and the biota act much like a nutrient "sponge" (a nutrient sink).
Laguna Cartagena sometimes experienced very dramatic events of P release and these dynamics were influenced by the hydrologic regime due to the abrupt periodic flooding and drying cycles imposed on it by agency management. The USFWS abruptly lowered the water level during the summer of 2010, leading to oxidation of the lagoon's exposed soils caused by exposure to the air and increased retention of P. This oxidation can result in the conversion of organic P to inorganic P, which can be subsequently released into water column [28].
In July 2010, USFWS re-installed all the flashboards to the top of the outlet water control structure, leading to a relatively long residence time of water accumulating behind the structure. Heavy rains fell in August and September of 2010, leading to an elevating water level (without outflow from the lagoon), and the rehydration of the lagoon's exposed dry organic soils, releasing P. High rainfall events and increasing water levels resulted in resuspension of bottom sediments, high turbidity levels and a substantial increase in TP concentrations of the lagoon water column [29]. High water levels lead to reducing conditions, where oxygen is absent. During Phase One (2010-2011), the lagoon exhibited hypoxic dissolved oxygen concentrations uniformly below 1 mg/L (ppm) that forced P release from sediments [30], forcing a greater flux of P from the organic matter, resulting in internal eutrophication [16,27,31,32].
In contrast, in Phase Two (2013-2014) the lagoon's P dynamic behavior was primarily influenced by internal biological and physicochemical processes within the lagoon. During this period Laguna Cartagena acted more often as a source of SRP and TP. The increased P release to the water column observed in Phase Two is likely the result of substrate disturbance (sediment and soil) and/or increased rates of plant decomposition in response to management interventions during the summers of 2012 and 2013; that reflect the transformation of sediment binding sites from imposed environmental variations in either oxidation or drying of the sediment, causing dramatic P release. Phosphorus release depends on the concentrations and distribution of P in the sediments, the degree of saturation of exchangeable phosphorus, the intensity of biological and chemical processes taking place at the water-sediment interphase and hydrological conditions [33]. During the post-intervention Phase Two (2013-2014), the lagoon exhibited levels of dissolved oxygen between 0.04 mg/L to 10.54 mg/L (ppm) [24,25], likely attributable to the initial removal of floating peat mats and floating vegetation, allowing greater water surface contact with the air and the penetration of light for photosynthesis by microalgae and submerged plants.
Several studies have demonstrated that under aerobic conditions or sediment aeration, P released due to microbial mineralization is transferred to the reductant-soluble, labile, and pore water pools [34].
Although sediment dredging is sometimes used in an attempt to reduce internal P-loading [31,[35][36][37], sediment removal can significantly disrupt other wetland processes because of physical disturbance and extensive drawdown of surface water [38]. Disturbance of P-enriched sediments has the potential to cause high phosphorus concentrations in the water column and low underwater irradiance due to sediment resuspension [39]. Sediments with high organic phosphorus promote the photo-release of inorganic phosphorus during resuspension of sediments exposed to simulated sunlight irradiation [40]. Aerobic sediments transforming to anoxic conditions release more P because of higher microbial respiration rates. Moreover, upon refilling, as the lagoon stage increased, more soils were inundated, thus releasing more P. In this study, high turbidity was coincident with increasing TP concentrations, suggesting that the lagoon may have experienced an increase in the amount of mineral turbidity in its water column, again due to greater sediment resuspension and exposure of new sediments to the water column [4].
Thus, the P retention capacity of soils tends to diminish on reflooding after drying, compared with continually flooded soils [27,29,41,42]. Thus, as Laguna Cartagena was drained there were concurrent changes in decomposition and the state of nutrients, making phosphorus more soluble, facilitating its export. Prescribed burning, employed to reduce organic matter and detritus, also converts varying proportions of the biomass and soil organic matter to ash and thus mineralizes nutrients [38,43]. Studies elsewhere, for example, in the Orlando Easterly Wetland (OEW), showed a decrease in SRP concentrations but an increase in the dissolved organic phosphorus (DOP) and particulate phosphorus (PP) resulting from a burn in the OEW [38], consistent with our observations here.

Conclusions
This study investigated the effects of intensified management interventions including excavation and removal of sediment and vegetation, draining, and burning on soluble reactive phosphorus (SRP) and total phosphorus (TP) concentrations of lagoon water, focusing on SRP and TP input and export during September 2010-September 2011 (Phase One) and October 2013-November 2014 (Phase Two), and including the roles of precipitation and water level changes on P sink and source dynamics. The results of water and soil/sediment analyses showed that Laguna Cartagena is a hypereutrophic and anoxic ecosystem that acted various times as both a sink and source of TP and SRP with internal phosphorus loading (internal eutrophication) from the lagoon's own sediments, causing higher phosphorus concentrations in lagoon water, downstream to other ecosystems.
Release of SRP and TP at the lagoon's outlet during Phase Two was substantially higher than during Phase One. We found no significant differences between P concentrations at the inlet canal for Phase One versus Phase Two. The lagoon also showed depleted N concentrations in the water column. During Phase One, P dynamics within the lagoon were primarily influenced by rainfall (one and two weeks prior sample collection) and water flows to the lagoon. In Phase Two the lagoon acted more often as a source of SRP and TP and this dynamic behavior was primarily influenced by internal biological and physiochemical processes within the lagoon. The USFWS Laguna Cartagena Conservation Plan emphasizes wildlife diversity and habitat restoration, but for this to be achieved the prevention and control of nonpoint P pollution is required, including the control of internal eutrophication. While nutrient and pollution inputs to the lagoon are beyond the control of the USFWS, the agency can nonetheless provide optimal wildlife habitats (keeping vegetative cover to about 30% instead of 90%) through the use of a common floating surface dredge. This would avoid disturbing bottom sediments, thus preventing excessive release of P to the lagoon water column and downstream to other water bodies and the Caribbean Sea. We recommend allowing normal (moderate) seasonal stage level fluctuations, but the complete cessation of draining, dredging, burning, excavations, and other interventions that can alter or disturb sediments.